Reviewers' comments:
Reviewer #1: The review of "Channel Selection Using Information Content Analysis: A Case Study of CO2 Retrieval From Near Infrared Measurements" (JQSRT-D-09-00251) for the Journal of Quantitative Spectroscopy and Radiative Transfer.
We are grateful to the reviewer for the careful reading and the many insightful comments. We have incorporated most of the reviewer’s suggestions. Our responses are in italics.
Main Comments:
The paper is concise and well written. The main concern with the paper is its focus that the general microwindow technique described in the paper is new. Information content based microwindow selection algorithms for a large number of geophysical parameters (not just temperature and water vapor) are well established in the scientific community (e.g. MIPAS (von Clarmann and Georg Echle (1998); Dudia et al., (2002)). Some of these microwindow selection algorithms are even more advance then the one outlined by the authors in this paper. For example, Worden et al., (2004) include estimated systematic errors (e.g. spectroscopy) in the TES microwindow selection process. Therefore, comments in the paper such as, "this technique is general and equally applicable to the retrieval of other geophysical variables (e.g. temperature or CH4), or modified for other instruments, such as AIRS or TES.", does not make a lot of sense as instruments like TES already use information based
microwindow selection tools that are more advance is some aspects than the one described in this paper. The authors need to better reference previous work, and thus, remove the assertion that the microwindow selection method described in the paper is new.
The new work this paper provides is the selection of CO2 microwindows in the NIR (using a standard information based approach). This could be used as reference for the scientific community performing NIR retrievals of CO2, however, the author's should also expand in more detail why microwindows are important. The negative side of using microwindows instead of the whole spectrum is that redundant information in the whole spectra could help reduce random instrument errors in the retrieval. However, there are some advantages for using carefully selected microwindows. The author's mention that microwindows greatly increase the speed of the retrievals. Microwindows can be used to avoid systematic errors (e.g. known spectroscopy errors, interfering species, etc.). If systematic error estimates were also included in the analysis the microwindows derived in this paper, the microwindows would be of greater benefit to the CO2 retrieval community.
The authors agree with the reviewer's assertion that microwindow selection based on information theoretic analysis is not new. However, that is not the assertion we make in this paper. Our new contribution is the selection of individual channels using information theory. Only one paper (Clarmann and Echle, 1998) even addresses the possibility of retrievals using carefully selected individual channels. However, they do not go into any details and instead focus on microwindow selection. Based on our analysis in this paper, we still believe that our technique is general and applicable to the retrieval of any geophysical quantity and from several existing and future instruments. TES does use information theory-based microwindow selection. However, that is different from the kind of retrieval we propose. At the very least, it should be worthwhile to compare retrievals using the two approaches since it would provide a quantitative estimate of the pros and cons of each technique.
As for the estimates of systematic errors, we do have a brief description of the errors in CO2 retrievals caused by uncertainties in temperature, surface pressure and water vapor. This is based on an analysis of synthetic data and we felt that it did not warrant more in-depth studies. However, we do have access to GOSAT measurements and will be exploring this aspect in much greater detail in a subsequent paper.
Minor Comments
1. Pg. 3, Par. 2 : Add some more references for the background CO2 statements
We have added more background CO2 statements in the third paragraph of page 3.
2. Pg. 3. Par. 2: The authors are well aware the GOSAT provides measurements in both the TIR and NIR as they mention it later in the paper. However, the way the last two sentences in the paragraph read it implies that GOSAT measures in the NIR and AIRS and TES measure in the TIR. Just add something like, ".(GOSAT) [3-4] is providing space-based measurements in both the near infrared (NIR) and thermal infrared (TIR)."
We have made changes as below.
“The JAXA Greenhouse gases Observing SATellite (GOSAT) [3-4] is providing space-based measurements in both the near infrared (NIR) and thermal infrared (TIR).”
3. Pg 4, Par. 1: "Both AIRS and TES provide measurements of the thermal infrared (TIR) CO2 band at 15 um." A minor point, but I think in addition to microwindows in the 15 um band the TES Science also uses microwindows in the laser bands for their CO2 retrievals.
We agree with the editor. The following sentence has been added:
“In addition, TES also uses two laser bands at 967-990 and 1070-1117 cm-1 for their CO2 retrieval [11].”
4. Pg 4. Par 3: It is important that the authors mentioned, "Further, it is very hard to eliminate biases due to correlations between parameters". As mentioned in the major comments section the authors need to expand the positive and negatives for using microwindows and to list potential biases (systematic errors).
We agree with the editor and have added more discussion as follows after that sentence.
“Clarmann and Echle (1998) discussed the selection of the optimum microwindows with respect to their associated retrieval errors [13]. The sources of retrieval errors are random errors of the measurement, and errors of the forward model and its input parameters. One goal of channel selection is to make an optimum trade-off between random measurement errors and systematic errors. Adding more channels usually decreases random measurement errors but increases the sysmetatic errors.”
5. Pg. 4, Par 4: As mentioned in the major comments section, there are many instrument science teams (MIPAS, TES, etc) that use information content analysis for microwindow selection for a wide range of geophysical parameters, not just temperature and water as stated here.
We have changed the beginning of this paragraph as follows:
“There has been some previous work on the optimization of retrievals from high spectral resolution measurements using information content analysis. Most of the earlier work has focused on choosing channels for retrieving temperature, humidity and other geophysical parameters [13–15]. However, none of these studies considered the selection of CO2 channels in the NIR.”
6. Pg. 5, Par 1: Change "ARIS" to "AIRS"
We have made the correction.
7. Pg. 12, Par 1: As mentioned above, "the same method could be applied to perform retrievals for measurements by other instruments such as AIRS, TES, or GOSAT." This does not make a lot of sense since instruments like TES already use general techniques like the one described in the paper. Essentially, most scientific teams using an optimal estimate type of retrieval approach have developed microwindows selection tools using information content.
We have made changes according to the reviewer’s comments.
“We intend to apply this method to CO2 retrievals from GOSAT measurements.”
References:
Anu Dudhia, Victoria L. Jay, and Clive D. Rodgers, "Microwindow Selection for High-Spectral-Resolution Sounders," Appl. Opt. 41, 3665-3673 (2002).
Thomas von Clarmann and Georg Echle, "Selection of Optimized Microwindows for Atmospheric Spectroscopy," Appl. Opt. 37, 7661-7669 (1998.)
J. Worden, S. S. Kulawik, M. Shephard, S. Clough, H. Worden, K. Bowman, and A. Goldman, "Predicted errors of Tropospheric Emission Spectrometer nadir retrievals from spectral window selection," J. Geophys. Res., vol. 109, no. D09308, May 2004.